The proposed research is designed to gather evidence for the validity of measures of caretaker "burden" or more broadly "impact". Since there is no "gold standard" criterion that can be used to validate such measures our aims are designed to develop and test validity through multiple approaches. Specifically, we use expert judgement to assess content validity, factor analysis to test notions about the structure of items measuring theoretically identified constructs, convergent validity by determining whether measures correlate with clinical judgement of the same domain, and construct validity by testing hypotheses concerning how our scales should correlate with other scales if they measure what we say. Of these approaches the last is central. To achieve it we must set forth a theoretical approach that generates hypotheses to test. We do this by developing a theoretical approach that links research on burden to research on the stress process. This conceptual innovation allows us to test hypotheses about our measures and to generate evidence concerning construct validity. To achieve our aims we propose to collect data from three samples: Alliance for the Mentally Ill (N=250), New York Hospital (N=100), and Saint Lukes/Roosevelt (N=150). These samples were selected because they vary in the extent of experience the relative will gave had as a caretaker, the degree of identification with the role as evidenced by membership in AMI, and in sociodemographic characteristics. This variety will allow us to assess the applicability of our measures across different kinds of subjects. Achieving our specific aims will contribute to the development of knowledge about the demands of the caretaking role. Moreover, linking our research to the stress process allows the possibility of determining how best to intervene to reduce the stress of caring for a mentally ill relative.